J. E.tJmolJiol. 9(2):135-155 Winter 1989

THE DEVELOPMENT OF A MUTUALISTIC RELATIONSHIP BETWEEN HUMANS AND MAYPOPS (PASSIFLORA INCARNATA L.) IN THE SOUTHEASTERN UNITED STATES

KRISTEN JOHNSON GREMILLION Departmmt of A'It11ropology 301 All/nUl; Buildi1lg, Campus Box 3115 University of North Carolina Chapel Hill, NC 27599-3115

ABSTRACT.-Archaeological evidence indicates that use of PassifIora incorntlta L. (maypops or passionnower) increased prehistorically from its earliest appearance in the Late Archaic. The intensification of the relationship between humans and this species can be understood with reference to ty,'o important components of human/ plant mutualism, seed dispersal and environmental modification. Increased usc of maypops is shown to have occurred along with the spread of plant husbandry and associated disturbance of existing vegetation. In addition, an extension of its range in modern times has apparently occurred despite its lack of importance as a food plant among Euroamerican populations. These findings support the notion that both anthropogenic habitat extention and seed dispersal conlribute to the development of human/plant mutualism.

RESUMEN.-la evidencia arqueolagica indica que el uso de Ptmif1t:Jra im:arrwta L (pasionaria) aumento prehistoncamente a partir de 50 primera aparid6n en el An:aico Tardio. La intensificaciOn de las relaciones entre los humanos y esta especie puede entenderse con referenda a dos componentes importantcs del mutualismo entre seres humanos y plantas: 1.1 dispersion de las semillas )' la modiflCaci6n ambiental. Se muestra que un mayor usa de 1.1 pasionaria sc dio junto con 1.1 expansiOn de 1.1 agricultura y 1.1 perturbacion concomitante de 1.1 vegetad6n natural. Parece, adem,h, haber ocurrido una extension de su distribUCKlO en tiempos recientes, a pesar de su falta de importancia como fuente de alimento para 1.1 poblaci6n euroamericana. Estos resultados apoyan la idea de que tanto 1.1 expansion del h~bitat antropogcnico como la dispersion de las semillas contribuyen .11 desarrollo del mutualismo entre humanos y plantas.

RESUME.-L'cvidence archeologique indique qu(' I'usage de Passif/ortl irrcamata L. (la passiflorc) s'est aumentee prchistoriquemente de sa premiere apparition dans l'Archaique Dcrnier. On peut comprendr(' I'aumentation de ce rapport enlr(' les humains et cetle esp~e avec la r(:ference a deux composants importantes du mutualisme enlre les humains elles plantes: 1.1 dispersion des graines et Ja modifica­ tion de I'environs. On vois que l'utilisation aumentee de 1.1 passinore s'est passee tout ensemble avec I'extension de I'agriculture el du derangement de la vegetation naturale. En plus, I'extension de sa portee dans les temps modemes s'est passk cvidemment malgn'; Ie manque d'importance comme une plante comestible entre 1.1 population euroamericaine. Ces conclusions soutienncnl I'idee que I'extension de I'environs et aussi la dispersion des graines ont contribue au dcvcloppement du mutualisme entre les humains ct les plantes.

INTRODUCTION Although students of prehistory have often taken an interest in human-plant relationships, this interest has most often centered around the origin of agri- 136 GREMILLION Vol. 9, No.2 cultural systems and its cultural and demographic consequences. Less attention has been given to relationships between human groups and plant species of minor economic importance. Recently ethnobotanists concerned with eastern North American prehistory (e.g. Asch and Asch 1985, Fritz 1986, B. Smith 1987) have given increasing attention to indigenous domesticates such as chenopod, sump­ weed, and sunflower and have recognized a number of minor commensal crops. Maypops (Passiflora incarnata L.) has been recognized as a weed crop on the basis of its common occurrence archaeologically and its preference for anthropogenic habitats (Yarnell 1987). Perhaps because it was a minor food plant rather than a staple, however, no attempt has been made to examine archaeological evidence for changing use of maypops over time or to discuss in any detail the relation­ ships between this species and aboriginal North American groups. In addition to compiling and discussing archaeological and ethnohistoric evidence for use of maypops and summarizing temporal patterns, this paper describes the geographic range of the species and relevant ecological characteristics. After reviewing ethnohistoric and archaeological evidence for increasing aboriginal use of maypops, I describe how certain propositions derived from Rindos (1984) about the roles of seed dispersal and environmental disturbance in the development of human-plant mutualism can be applied to human relationships with this species. I then compare the archaeological evidence of the use of maypops, vegetational change, and aboriginal relationships with other plant species with these propositions. Using these data, long-term changes in the use of maypops in the Southeast can be explained in part with reference to its mutualistic relationship with human populations. Probable range extension of the species in modern times is also discussed in the context of this model of domestication.

SPECIES CHARACTERISTICS

The Passifloraceae, of which Passiflora incarnata L. (maypops) is a member, is a predominantly tropical family with some 23 genera, four of which are found exclusively in the New World. Passiflora is a pantropic genus of some 400-500 species, most of which occur in the American tropics (Hutchinson 1967, Martin and Nakasone 1970).1 The native range of maypops covers much of the eastern United States from Virginia to southern Illinois and southeastern Kansas south to Florida and Texas (Killip 1938, 1960). P. lutea L. is the only other species in the Passifloraceae that occurs naturally in the continental temperate United States. Maypops is a perennial vine with showy flowers and an edible fleshy fruit. It is frequently an aggressive colonizer of vacant lots, railroad beds, old fields, and cultivated fields in the Southeast. Maypops may also be found along road­ sides and fencerows or in thickets (Radford et al. 1968:734). Its distribution in these kinds of habitats indicates a preference for disturbed soils and exposure to full sunlight. Judging by the species' present-day preference for anthropogenic habitats, it presumably was restricted to relatively open ground beneath canopy gaps (including stream margins, windthrow gaps, and other naturally disturbed patches in the landscape) before disruptive human activity became common within its range. Winter 1989 JOURNAL OF ETHNOBIOLOGY 137

Although detailed data on the ecology of maypops are unavailable, some aspects of the morphology of its flowers and fruits allow for inferences about its reproductive biology. Maypops flowers are quite showy; in fact, the species is sometimes grown today as an ornamental in the United States. Showy flowers are nearly always characteristic of animal pollinated plants. The maypops fruit is fleshy and sweet and the numerous seeds have woody walls. Thus the fruit is likely to be consumed by animals without resulting destruction of the seed. Given these aspects of fruit and seed morphology (fruit palatability and seed durability), mammal or bird dispersal is most likely.

MODERN DISTRIBUTION

The present-day distribution of maypops is presented in Figure 1. Range information was obtained from published sources where possible and from herbarium specimens (primarily from the University of North Carolina at Chapel Hill). Occurrences were noted by county when this information was available in published sources or from herbarium specimens (see Appendix for a list of references). In addition to the county collections shown in Figure I, published sources indicate that maypops is found throughout Louisiana, Tennessee and Georgia. In addition to most of the states in the Southeast maypops also occurs in southern Virginia, western West Virginia, western Kentucky, southern Illinois,

MILES o 100 200 300 400 I, I,', I I" I", ,

FIG. I.-Modern distribtltion of maypops. Each dot represents a record of collection for a COtlnty. 138 GREMILLION Vol. 9, No.2

southern Indiana, southern Missouri, eastern Nebraska, Oklahoma and Texas, and throughout Arkansas. Maypops occurs in southern Florida, but it shares its range there with several congenerics with Caribbean affinities (Avery and Loupe 1980). TIle single occurrence for Pennsylvania reported by Killip (1938) is not listed by Wherryet al. (1979) and has therefore been omitted from Figure 1. Apparently maypops has been introduced as an ornamental north of its present natural range from Virginia west to Missouri (Killip 1938). The western limit of the range of maypops correlates well with the western limits of the eastern deciduous forest region. This correlation is best explained as il similarity in moisture requirements between maypops and deciduous forest trees, whose growth westward is limited by low soil moisture in transilional grassland regions (Vankat 1979:132). It is unlikely that maypops' range is limited by some aspect of the environment dependent upon the presence of forest trees, since it is a vigorous colonizer of treeless areas and should be expected to thrive in them as long as conditions for growth and reproduction are met. The northern limit of the species' range is probably the result of an intolerance to cold andlor a photoperiodic requirement for flowering to occur. All collections of P. incarnata (with the exception of the report from Pennsylvania noted above) have been made within the area having an average frost-free period of 160 to 200 days (Espen­ shade 1982). There is no reason to suspect that maypops is limited to particular kinds of substrates, since it occurs on all soil types. It does prefer disturbed soils, and as a weed, it is heliophilic and opportunistic of disturbances that remove or alter existing vegetation. Tolerance of a relatively wide range of environmental condi­ tions is also characteristic of weeds (Baker 1974). 50 is a preference for anthrO"­ pogenic disturbances, the frequency and extent of which influence the abundance of maypops in many areas and may have a significant effect upon its geographic range as well.

PREHISTORIC AND EARLY HISTORIC DISTRIBUTION AND USE

Documentary euidence.-Two sources, the accounts of William 5trachey (Major 1849) and Robert Beverley (Wright 1947) are particularly informative about aboriginal relationships with maypops in the East. Both wrote of coastal Virginia, inhabited by Algonkian peoples. 5trachey's account of early seventeenth century Indian life reports that "they plant also the field apple, the maracock, a wyld fruict like a kind of pomegranett, which increaseth infinitlye, and ripens in August .. " (Major 1849:72). He also describes the plant and its growth habits at some length:

Here is a fruict by the naturells called a maracock; this groweth generally low, and cTeepeth in a manner amongst the corne (albeit I have seene yt, planted in a gardein within our fort, at James Towne, to spred and rise as high as the pale); yt is the bignes of a queen apple, and hath manie azurine or blew kamells, like as a pomegranet, and yt bloometh a most sweet and delicate flower, and yt is a good sommer cooling fruict, and in every field where the Indians plant their corne be cart-loads of them. (Major 1849:119) Winter 1989 JOURNAL OF ETHNOBIOLOGY 139

These comments indicate that maypops was abundant in Indian gardens. The latter comment even suggests that maypops and gardens were usually associated. Although supporting the idea of a symbiotic relationship between human groups (and their managed landscapes) and maypops, Strachey's account equivocates on the question of whether or not the species was planted. In the same sentence we are told both that the plant was wild and that it was planted by the Indians. Beverly's observations of the early eighteenth century in the same area indicate that maypops was a weed rather than a crop: The Maracock, which is the Fruit of what we call the Passion Flower, our Natives did not take the Pains to plant, having enough of it growing every where; tho' they eat it with a great deal of Pleasure; this Fruit is about the Size of a Pullet's Egg (Wright 1947:142). Perhaps these contradictory reports reveal a European disinclination to recognize unfamiliar types of plant management that seemed to elude a perceived dichotomy between "wild" and"domesticated." Alternatively, the nature of the human­ plant relationship involved may have changed between Strachey's time and Beverly's (although the earlier report also seems uncertain as to the status of maypops, assuming a connotation of the term "wild" that is similar to ours today). In any case, these two sources do indicate that maypops grew abundantly in Indian gardens, that the fruit was consumed, and that the plant was tolerated and possibly encouraged or planted in Historic times. Once planted in cultivated soil, such a persistent weed would have needed little encouragement to flourish. Its reported abundance in gardens is evidence of its success as a colonizer of human-disturbed habitats even without planting. Archaeological evidence.-Seeds of maypops are common components of paleo­ ethnobotanical assemblages in the Southeast and are found occasionally in adja­ cent regions. Incorporation of its carbonized seeds into archaeological deposits is unlikely to have occurred without human manipulation of the fruits, which otherwise would decompose and disseminate their propagules far from hearth fires. Thus their archaeological occurrence may be considered strong evidence of use of the fruit by aboriginal populations. The plausibility of consumption of the fruits is further supported by the modern-day use and cultivation of Passiflora throughout the tropics (Popenoe 1974:242; Hedrick 1972:409-411). In order to explore this question of changing use over time, paleoethnobotanical data were assembled from a number of published and unpublished sources. For the Archaic and Woodland periods, original sources were consulted in conjunc­ tion with Yarnell and Black's (1985) compilation of data. For the Mississippian/ Protohistoric and Historic periods, various reports were reviewed. All sources are listed in Table 1 along with the sites and components from which maypops seed counts were derived. For each component at each site, absolute quantity of maypops seeds recovered was obtained as well as total number of fleshy fruit seeds. Although acquisition of accurate totals was a priority, it must be recognized that difficulties are often encountered in synthesizing data of this kind and that differences between researchers in recovery, processing, analysis and quantifica­ tion render comparisons approximate rather than precise. However, it is felt that ~ TABLE I.-Sites with maypops seeds. ~

Period Site Name Site Number State County Temporal Context Reference

Late/Term. Archaic Iddins 40LD38 TN Loudon 1800-1300 BC Chapman & Shea 1981 (3500-800 BC) 1FR310 AL Franklin Perry Caddell 1982a Early Woodland 9CK(DOT)7 GA Cherokee Kellogg (800 BC- 565±75 BC- Bowen 1982 AD 200) 700±75 AD Middle Woodland 1FR310 AL Franklin Lick Creek Caddell 1982a BC~- CJ (100 Icehouse :;:0 tT1 AD 550) Bottom 40MR23 TN Monroe Connestee/Candy Creek/ Chapman & Shea 1981 ~ AD 200-500 L4 Eoff I 4OCF32 TN Coffee McFarland Faulkner 1977 aL4 Late Z Woodland Mason 40FR8 TN Franklin Mason!AD 600-1000 McMahan 1983 (AD 550-1000) 1Grlx1 AL Greene Middle Miller IIIAD 900 Caddell 1982c 16WC5 LA W. Carroll Coles Creek Shea 1979 1JE31 AL Jefferson West Jefferson Scarry 1980 1JE33 AL Jefferson West Jefferson Scarry 1980 Miss./Protoh. Moundville AL Hale Moundville I1AD 900-1000 Scarry 1986 (AD 1000- Martin Farm 40MR20 TN Monroe Miss. I and II Schroedl et al. 1985 1550) Lubbub Creek locality 1PI33, 83, 85 AL Pickens Summerville I-III Caddell 1983 Beaverdam <: Creek 9EB85 GA Elbert Savannah II Gardner 1984 ~ Gordon ....\0 Z Mounds 22JE501 MS Jefferson Late Coles Creek- Shea 1984 9 Plaquemines N TABLE i.-Sites with maypops seeds. (continued) ~ ro.... ~ Period Site Name Site Number State County Temporal Context Reference \0 00 \0

Cemoche- chobee 9Cla62 GA Clay Mississippian Dunn 1982 Toqua 4OMR6 TN Monroe Dallas Shea et ale 1986 38AN8 SC Anderson Late Miss. Gardner 1982 44HA22 VA Halifax Protohistoric Trinkley 1976

Yarborough 22CL814 MS Clay Protohistoric Caddell 1982b "- Wall 310R11 NC Orange Protohistoric Gremillion 1987b c:0 :;0 Hist. Little Egypt 9MU102 GA Murray Barnett/AD 1550-1700 Hally 1981 Z

(AD 1550- >~ 1800) Tanasee 4OMR62 TN Monroe Cherokee Schroedl & Shea 1986 0 "T1 Chota 40MR2 TN Monroe Cherokee/AD 1746-1774 Schroedl & Shea 1986 tT1 Lamar Village GP-HK-08 GA Hancock early 1600s Gardner 1985 Mialoquo 40MR3 TN Monroe Cherokee Russ & Chapman 1983 ~ 0 Upper ao::t Saratown 31SKla NC Stokes late 1600s Wilson 1977 ~ 0 Mitchum 31Ch452 NC Chatham 1600s Gremillion 1987b C) Fredricks 310R231 NC Orange late 1600s Gremillion 1987a .-< 1 Misc. Spoonbill 41WD109 TX Wood AD 750-1350 Perttula & Bruseth 1983 Calahan- 23MI171 MO Mississippi AD 1300-1500 Byrd & Neuman 1978 Thompson Towasahgy 23MI2 MO Mississippi L. Mississippi? Cottier 1971 Nuyaka ViI. 1HB3 AL Tallapoosa AD 1000-1200 Crawford 1975 Spiro 34LE46 OK LeFlore AD 1000-1400 Fritz 1982, 1984 Duncan Tract 4OTR27 TN Trousdale prehistoric Shea 1983 22HI500 MS Hinds AD 1200-1400 Cutler & Blake 1976

~ ~ lUsed only for constructing Figure 3 (not included in quantification.) ~ 142 GREMILLION Vol. 9, No.2 the database is large enough to make comparisons among time periods useful for assessing change. Quantitative data were drawn only from sites in the Southeastern states (here including Alabama, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, and Virginia). Two quantitative methods were employed in arriving at a measure of the frequency of use of maypops for each time period. The first involved calculating the proportional contribution of maypops seeds to total numbers of fleshy fruit seeds2 for each site and each time period based upon sources listed in Table 1. The same operation was performed on data presented in Chapman and Shea (1981) to create a sequence showing changes in the archaeological occurrence of maypops in a particular region, the Tellico Reservoir area of the lower Little Tennessee River valley. For sources listing fragments, numbers of whole seeds were added to half the number of fragments in order to estimate numbers of whole seeds. The other method used was calculation of ubiquity as the percentage of sites of each time period from which maypops seeds were recovered. The same dataset was used as for calculation of seed percentages. However, the total number of sites from which the percentage containing maypops was calculated is the total number of sites for each time period for which a report on paleoethnobotanical remains was reviewed by the author. These data were obtained by consulting Yarnell and Black (1985) for the Archaic and Woodland periods and by reviewing published and unpublished reports on plant remains for the Mississippian, Protohistoric and Historic periods. Thus ubiquity values are in part a function of the quality and quantity of research done in various regions and of the completeness of the author's survey of available data. Since ubiquity is less subject to biases produced by differential preservation than are manipulations based upon seed quantities, this measure is probably a more reliable indicator of change. Table 2 presents the proportional contribution of maypops seeds to the total number of fleshy fruit seeds recovered from Southeastern archaeological sites through time. A general trend toward comprising a greater percentage of fleshy fruit seeds is apparent. As a group, the Late Archaic through Late Woodland periods produce low maypops percentages compared to the Mississippian! Protohistoric and Historic periods, although these values do not increase steadily over time. The Late Woodland percentage is lower than either the Late Archaic or Early and Middle Woodland. The relatively small number of sites used in calculations for these earlier time periods are likely to be responsible for this devia­ tion. Nonetheless, there seems to be a dichotomy between pre-Mississippian and post-Mississippian representation of maypops seeds in archaeological deposits that correlates with the beginnings of the preeminence of maize as a crop plant in the East. An even clearer pattern is obtained by comparing quantities of seeds recovered from different time periods within a more restricted geographical area. Data from Chapman and Shea (1981) from sites in the Tellico Reservoir were used for this purpose. Quantities obtained for each time period and representation of maypops as percentage of fleshy fruit seeds for the Tellico Reservoir are given in Table 3. In this case, there is clearly an increase in the representation of maypops Winter 1989 JOURNAL OF ETHNOBIOLOGY 143

TABLE 2.-Maypops seed counts and percentages.

No. of No. of % Period Site Maypops Aeshy Maypops Seeds Fruit Seeds Seeds

Late Archaic 4OLD38 1 284 0.5 IFR310 5 5 100.0 Total 6 289 2.1

Early/Middle Woodland 9CK(DOn7 1 3 33.3 IFR310 5 6 83.3 40MR23 4 46 8.7 4OCF32 2 146 1.4 Total 12 201 6.0

Late Woodland 4OFR8 3 105 2.9 IGrlx1 1 6 16.7 16WC5 1 69 1.4 ljE31 1 SO 2.0 IJE33 2 332 0.6 Total 8 562 1.4

Miss.lProtoh. Moundville 44 1304 11.0 40MR20 13 82 15.9 Lubbub Creek 7 37 18.9 9E885 166 234 70.9 22JE501 1 2 SO.O 9C1a62 2 261 0.8 38AN8 12 SO 24.0 44HA22 3 40 7.5 22C1814 21 139 15.1 310Rll 10 39 25.6 40MR6 53 266 19.9 Total 332 2454 13.5

Historic 9MUI02 11 511 2.2 40MR62 6 29 20.? 40MR2 53 103 51.5 GP-HK-DB 28 29 %.6 40MR3 3 7 42.9 31SKla 15 249 6.0 31CH452 5 20 25.0 310R231 64 271 23.6 Total 185 1219 15.2 144 GREMILLION Vol. 9, No.2

TABLE 3.-Tellico Reservoir sites, maypops percentages by period.

No. Maypops No. Fleshy Fruit 0/0 Maypops Seeds1 Seeds1 Seeds

Late Archaic 1 284 0.4 Middle Woodland 4 39 10.3 Early Miss. 20 57 35.1 Late Miss. 43 872 49.4 Historic 118 359 32.9

1From Chapman and Shea (1981); estimates based on number of whole seeds + number of fragments/2. 2Excludes the -2752 persimmon seeds reported, most of which were recovered from a single structure. seeds from the Archaic to the Early Mississippian. Percentages are similar for the Early Mississippian, Late Mississippian, and Historic periods. A general pat­ tern of increasing relative quantities from Archaic to Woodland and from Wood­ land to Mississippian and Historic periods is evident. For the Southeast as a whole, calculation of ubiquity of maypops as per­ centage of sites from which it was recovered per time period produces stronger evidence of increasing use. These ubiquity values are displayed graphically in Figure 2. These data show a clear and steady increase in the likelihood of maypops being recovered from a site. The time periods can be roughly grouped into classes according to these ubiquity values. The first class (0%) includes the Early and Middle Archaic. The second (Late Archaic, Terminal Archaic and Early Woodland) % has values near 7 • The third class includes the Middle and Late Woodland, both with values of 16.7%. The fourth and fifth classes include the Mississippian/ Protohistoric and Historic periods, respectively. There is a strong positive cor­ relation between lateness in the temporal sequence (and, presumably, the extent and intensity of anthropogenic disturbance) and the representation of maypops in the archaeological record as determined by ubiquity. The geographical range of maypops during different time periods might also be used to assess its importance as a food plant in the Southeast. A range extension would be an expected consequence of husbandry of the species or simply of an increasingly close relationship with human populations. Figure 3 presents archaeological occurrences of maypops seeds constructed using sources listed in Table 1. Unfortunately, the archaeological record of the distribution of maypops is heavily biased by the nonrandom geographical clustering of research efforts. More importantly, the archaeological record is not extensive enough to provide a realistic depiction of the entire range of the species during different time periods. However, some aspects of the premodern range of maypops seem not to be merely products of sampling bias. In contrast to the modern range of the species (Fig. 1), there is so far no archaeological evidence of maypops from Kentucky, Illinois, or Arkansas. West-central Illinois has yielded a particularly large and well-researched body of paleoethnobotanical data (e.g. Asch and Asch 1985) and paleoethnobotanical research has been actively undertaken in both Winter 1989 JOURNAL OF ETHNOBIOLOGY 145

70

60

50

0) 40 0;;.! '0 it 30

20

10

EA MA LA TA EW MW LW M/PH H Period

FIG. 2.-Archaeological ubiquity of maypops. The vertical axis indicates the percentage of total sites reporting plant remains that also reported maypops seeds. E = Early, M = Middle, L = Late, A = Archaic, W = Woodland, M/PH = Mississippian Protohistoric, H = Historic. Total numbers of sites for each time period are: EA, 8; MA, 4; LA, 14; TA, 13; EW, 13; MW, 18; LW, 30; M/PH, 32; H,12.

Arkansas (Fritz 1986) and Kentucky (Watson 1985). Thus it is apparent that maypops either did not occur in these states in the prehistoric and early Historic periods or that it was not manipulated by people in such a way as to ensure its incorporation into the archaeological record.

IMPLICATIONS OF DISTRffiUTION AND USE PATTERNS

Having established a temporal pattern of increasing use of maypops in the archaeological record, it remains to be seen what hypotheses can be formulated to explain this pattern. Examination of evidence for the development of human­ plant mutualism can help to define the ecological relationships between aboriginal populations and this species and how it may have changed over long periods of time. The hypotheses generated must be consistent with the observation that a general trend to increased use (and presumably abundance) of maypops in 146 GREMILLION Vol. 9, No.2

MILES o 100 200 300 400 I, I, I" ,,' I" ,

FIG. 3.-Archaeological occurrences of maypops. Each dot represents one pre­ AD 1800 aboriginal site. The line indicates the approximate limits of maypops' modern range.

aboriginal contexts seems to have been followed by an extension of range in more recent times that coincides with its decline in importance as a food plant. Rindos (1984:143) has developed a model of domestication as a "coevolu­ tionary process in which any given taxon diverges from an original gene pool and establishes a symbiotic protection and dispersal relationship with the animal feeding upon it." Domestication is seen as a process in which both human and plant populations undergo change as a consequence of the increased fitness brought about by this relationship. If natural selection is the mechanism by which such a symbiotic relationship develops, an increase in the fitness of both symbionts is expected. Human populations benefit from these relationships through the energetic contributions of the plants involved, usually via their consumption. Two important components of human-plant relationships that enhance the fitness of the plant species involved and thus contribute to the development of mutualism are dispersal of propagules and environmental modification that increases the potential range of the plant.

Dispersal. - Animals may act as either generalized or specialized dispersal agents for plants. Generalized dispersal relationships are opportunistic whereas special­ ized dispersal usually results in obligate symbiosis and the evolution of mor­ phological traits in the plant that attract the dispersal agent and prevent effective Winter 1989 JOURNAL OF ETHNOBIOLOGY 147 dispersal by other animal species. Both kinds of dispersal relationship may act to stimulate the development of human/plant mutualism, but in general it is specialized dispersal relationships that characterize the more fully developed phases of the domestication process (Rindos 1984:112-120). Based upon its presumed dispersal relationship with humans, several char­ acteristics of maypops indicate the likelihood of a mutualistic relationship between these two species. The maypops fruit is fleshy and edible, and its seeds have a durable woody wall. In fact, germination of seeds of some PassifLora species is thought to be enhanced by passage through the human digestive tract (Hladik and Hladik 1967, cited in Rindos 1984:135). Presumably maypops seeds, which are small enough to be comfortably ingested by a human being, also retain viability thereafter. These characteristics argue for a general pattern of animal dispersal in the species and provide weak evidence in favor of the assertion that a disper­ sal relationship between maypops and humans existed in the aboriginal Southeast, whether or not the seeds were intentionally planted. Stronger evidence for human dispersal of maypops seeds is provided by the presence of 80 seeds found in association with a human burial in a context suggesting ingestion at the Windover site in Florida, dating to around 7200 BP (Newsom 1988). Ingestion of the seeds prehistorically can safely be assumed, if not the viability of the seeds after con­ sumption or the precise role of humans as dispersal agents. Was this dispersal relationship opportunistic or obligate? The characteristics of the maypops fruit today do not point to an obligate dispersal relationship with humans. The fruits, although relatively large, are small enough for other mammals to consume, and the seeds themselves are small enough and sturdy enough to pass through other types of mammalian digestive systems and retain viability. In contrast, fleshy fruit-bearing species that rely on a more or less obligate dispersal relationship with humans often have unusually large seeds that pre­ vent effective dispersal by smaller, non-human animals (Rindos 1984:117). Other correlates of this type of dispersal relationship include nutritional content and fruiting season. Species relying on specialized dispersal tend to be high in energetically expensive lipids and proteins, whereas opportunistically dispersed species contain the more common (and "cheaper") carbohydrates and water (Rindos 1984:116). Since the maypops fruit is high in water (89.6%) and carbo­ hydrates (7.9%) (Paul Gardner, pers. comm., 1987), its characteristics more closely approximate those of the typical opportunistically dispersed fruit. The fruiting seasons of plants with obligate dispersal agents tend to be long thus ensuring that the dispersal agent will find and consume sufficient quantities of the fruit (Rindos 1984:118); by this criterion maypops, which fruits in late summer and early fall in the Southeast, seems to be opportunistically dispersed. In general, maypops fits the general pattern of temperate-zone angiosperms in lacking a specialized dispersal agent (Rindos 1984). There is also no evidence that human populations had come to depend upon maypops as a food source, in contrast to what seems to have been the case with many starchy grain crops worldwide. Reliance on domesticates (Le., obligate symbiosis) to maintain population densities at newly-established levels is not likely to occur when only fleshy fruit-producing species are involved, since these are dietary supplements rather than staples (although such supplements may be 148 GREMILLION Vol. 9, No.2

important nutritionally, if not calorically). The coevolutionary relationship is more likely to have been opportunistic than obligate for both symbionts.

Environmental modifications.-Another aspect of the domestication process, namely the creation and maintenance of disturbed habitats, can also be applied to the development of a mutualistic relationship between humans and maypops. The role of anthropogenic disturbance in the development of plant husbandry has long been recognized (Anderson 1952) and recently reconsidered in light of new archaeological evidence (B. Smith 1987). Development of anthropogenic habitats is an important component of human/plant mutualism since it benefits certain species by increasing their reproductive potential near human habitations (Rindos 1984:140). Dumpheaps and middens are examples of such anthropogenic habitats (Anderson 1952, Dewar 1982) as are old fields and abandoned gardens. The latter are parts of an already existing agricultural ecosystem. Maypops is today well-adapted to such disturbed habitats and in fact seems to require them in order to establish large populations.

A case study.-Perhaps it is more profitable to assess the roles of dispersal and disturbance in the development of mutualism between humans and maypops in the wider context of human/plant relationships. If domesticatory systems are considered, a temporal sequence of increasing interdependence of humans and certain plants can be proposed for the Southeast. Although full discussion of such a hypothesis exceeds the scope of this paper, Chapmanet ale 's (1982) study .of environmental disturbance and plant husbandry in the Tellico Reservoir can be used to summarize the changes involved and how they might be related to the development of symbiosis between humans and maypops. The domesticates Lagenaria and Curcurbita are present by the Late Archaic (although Cucurbita at this time may have been only a protodomesticate). Maygrass (Phalaris caroliniana Walter), a small grain best classified as a quasi-cultigen (Yarnell and Black 1985), is also relatively abundant (Chapman and Shea 1981). The relative representation of taxa classified as " disturbance-favored" in the wood charcoal spectrum changed little from the Middle Archaic to the Late Archaic. Judging by the presence of the incipient cultigens during this period, initial processes of domestication were in place. The Late Archaic is also when maypops first appears in the archaeological record of the Tellico Reservoir. By the Woodland period, the percentage of disturbance-favored taxa had increased to about 35%-45% from the 20%-30% of the Late Archaic. Cultigen sump­ weed Iva annua var. macrocarpa), as well as large quantities of chenopod (Cheno­ podium sp.) first appear in the Early Woodland at Tellico. Knotweed (Polygonum sp.) was an important small grain crop in some parts of the East, but seems to have been of only minor importance in the Tellico Reservoir, occurring in small numbers during the Early Archaic and Middle Woodland. The earliest maize is from a Middle Woodland component (Chapman and Crites 1987), although it apparently did not become important until the Mississippian (Champman and Shea 1981). Thus, the crop plants bottle gourd, pepo, maygrass, sumpweed, chenopod, knotweed and maize were all present by the Middle Woodland in the Tellico Winter 1989 JOURNAL OF ETHNOBIOLOGY 149

Reservoir. Although archaeological evidence of Middle Woodland maypops is sparse for the Tellico Reservoir, the relative percentage of maypops seeds increased from the Late Archaic (Table 3). Presumably by this time domestication had become well-established and associated behaviors such as planting, cultivating, and other forms of management had come to characterize human relationships with the abovementioned plant species. Thus, maypops was one species parti­ cipating in a network of domesticatory relationships evolving between plants and local human populations during the Woodland period. The wood charcoal evidence for increased environmental disturbance probably reflects both inten­ tional disruption associated with clearing and planting as well as unintentional modification that allowed for the increase of populations of weeds and weed crops or quasi-domesticates, including maypops. Elsewhere in the Southeast, as well as in the Tellico Reservoir, the Missis­ sippian period saw the florescence of fully developed agricultural systems. In the late Mississippian IIdisturbance-favored" taxa make up roughly half of the wood charcoal analyzed, compared to about 25% for the early Mississippian. Early and late Mississippian periods produced similar percentages of maypops seeds, although late Mississippian values are somewhat lower. The Historic period shows a definite increase both in disturbance (as reflected in wood charcoal evidence) and in representation of maypops. In general, the sequence presented in Chapman et ale (1982) supports the proposed relationship between maypops and anthropogenic disturbance. Maypops occupies a niche in various kinds of anthropogenic habitats ranging from slightly disturbed to agricultural (that is, dominated by human management of plant life cycles). Although itself not involved in an obligate dispersal relation­ ship with humans, this species seems to have become dependant on anthro­ pogenic disturbance for maintenance of large populations. In fact, development of truly agricultural ecosystems seems to have been accompanied by an increase in human consumption of fruits of this species (and presumably its abundance near settlements as well). These generalizations hold whether or not maypops was intentionally planted at any time during this chronological sequence. Maypops, then, developed a mutualistic relationship with people that was facilitated by anthropogenic disturbance and dispersal, both processes that frequently provide a foundation for plant domestication. It may have been planted as well as harvested, but it is equally likely that maypops increased in abundance by invading gardens and was encouraged because of its usefulness. Because there is evidence that maypops developed a mutualistic, non-obligate relationship with humans and their agricultural and proto-agricultural ecosystems, should maypops be considered a domesticate? The relationship between humans and maypops certainly fits the definition of domestication used by Rindos (1984). However, this definition is broad enough to encompass virtually all human-plant feeding relationships and is therefore of little use in differentiating between types of mutualism of varying intensity. Rindos (1984) divides the continuum of domesticatory relationships into categories according to the human behaviors (such as planting, storage and protection) and characteristics of plant populations associated with each. The sequence from incidental to specialized to agricultural domestication reflects the increasingly 150 GREMILLION Vol. 9. No.2 obligate nature of human-plant mutualism and the greater importance of environ­ mental modification, dispersal relationships, and human management in fueling coevolution. Whether maypops should be considered as characteristically parti­ cipating in incidental, specialized, or agricultural domestication cannol be satis­ factorily resolved because its relationship with people is opportunistic. Although it is well adapted to agricultural systems, maypops is also at home in proto­ agricultural and non-agricultural systems. Rindos' model is useful in identifying ecological processes that contribute to the development of mulualistic relationships between human groups and useful weedy opportunists such as maypops. The notion that domestication is a pro­ cess rather than an event also makes intelligible the seemingly ambiguous status of species that do not fall into "wild" or "domesticated" categories but rather between these extremes of the continuum. On the other hand, application of the subcategories of domestication docs meet with problems when faced with a species that seems to fit almost equally well into systems that might be char­ acterized as incidental, specialized or agricultural. Perhaps for the time being, a term such as Yarnell's "weed crop" or "quasi-cultigen" (1987) is best for species that seem to be intermediate between cultigens and "wild" (i.e., utilized but not managed) plant resources.

RANGE EXTENSIONS

The present-day distribution of maypops reflects an apparent extension com­ pared to its early Historic range (Fig. 3). At first glance this situation seems to be incompatible with the hypotheses generated about the roles of dispersal and anthropogenic disturbance in the development of human/maypops relationships in the Southeast. With the demographic decline of aboriginal population and subsequent replacement by European populations with different dietary and agricultural traditions, the species might be expected to experience a decline in abundance along with its symbiont. Its range also should be expected to con­ tract. But the apparent extension is a sign of the ecological flexibility of maypops and its lack of an obligate relationship with specific types of anthropogenic disturbance. The most likely explanations for this range extension are in fact compatible with Rindos' model and the hypotheses presented above. Although the feeding practices that are often crucial components of dispersal relationships were lost with the substitution of Euroamerican for Native American populations (and cultural systems), they have been replaced by the cultural practice of raising ornamentals. Extension of the range of maypops into Indiana, Ohio, Kentucky, and Illinois may be a result of the attempts of horticulturalists to raise maypops near their homes because of their showy flowers. The second component of this model of developing mutualism, anthropogenic disturbance, has increased since Euroamerican settlement. An extension of the range of any colonizing species should be expected under such conditions. Winter 1989 JOURNAL OF ETHNOBIOLOGY 151

SUMMARY

I have presented archaeological evidence indicating that mutualism between maypops and human groups developed along with plant husbandry based chiefly upon other plant species. I hypothesize that the symbiotic relationship between maypops and human groups (or maypops and human cultural systems) reached a stable level in which a mutually beneficial (Le., fitness-enhancing) relationship was maintained without the application of intentional behaviors such as planting, seed storage, and cultivation. In any case, I have shown that the patterns of aboriginal use of maypops can be explained in part by examining the evidence with reference to two important aspects of human-plant mutualism, seed dis­ persal and environmental modification. A modern extension of the geographic range of the species can also be explained simply with reference to these processes. This study has illustrated the applicability of theoretical models of plant domestication (e.g. Rindos 1984) to paleoethnobotanical data and has also explored some of the problems encountered therein. Further study of species like maypops that fall outside of traditionally accepted categories such as "wild" and "domesticated" will be a necessary prerequisite to an adequate understanding of the full range of human-plant relationships.

APPENDIX

The following sources were consulted for information used in constructing the modern distribution map for maypops (Fig. 1): Allen et al. 1975; Avery and Loope 1980; Braun 1943; Brown 1972; Brown and Brown 1984; Clewell 1985; Deam 1940; Great Plains Flora Association 1977; Harvillet al. 1977; Kellerman and Werner 1894; Lakela 1965; Mellenger and Whipple 1981; Mohlenbrock and Ladd 1978; Radford et al. 1968; Sharpet al. 1960; Small 1913; E. Smith 1978; Steyermark 1962; Strausbaugh and Core 1978; Wherryet al. 1979. In addition to the Herbarium of the University of North Carolina at Chapel Hill, the Herbarium of the University of Alabama provided range information used in constructing Fig. 1.

ACKNOWLEDGEMENTS

Earlier drafts of this paper were greatly improved by the comments of M. Jean Black, Jared Diamond, Paul Gardner, Robert K. Peet, Gail Wagner, Richard A. Yarnell, and an anonymous reviewer. In addition, Paul Gardner kindly allowed access to the results of his nutritional content analysis of maypops, and Richard Yarnell allowed me to use his abstracted versions of reports on plant remains from various sites. Jimmy Massey of the Herbarium at the University of North Carolina, Chapel Hill provided valuable advice on acquiring plant distribution information. Earlier versions of this paper were prepared for Biology 143 taught by Robert K. Peet at the University of North Carolina, Chapel Hill and for the Southeastern Archaeological Conference held in Charleston in November of 1987. Responses to requests for information sent to herbaria at Northern Kentucky University and the University of Alabama are gratefully acknowledged. Full responsibility for this article's contents are the author's alone. 152 GREMILLION Vol. 9, No.2

NOTES

1A vernacular name for the genus, passionflower, derives from the Italian fior della passione, applied to the flower because of its symbolic representation of objects, persons and events associated with the crucifixion (Fernald 1950:1042). The term maypops, often applied to the species, refers to its fruit.

2In addition to Passiflora incarnata, the following taxa reported in the sources consulted have been included in the fleshy fnlit category: Anzelanchier, Celtis, Crataegus, Diospyros virginiana, Fraga ria, Gaylussacia, Gleditsia triacanthos, Passiflora lutea, Physalis, Prunus, Rhus, Rubus, Sambucus canadensis, Solanum, Vaccinium, Viburnum, Vitis. LITERATURE CITED

ALLEN, C.M., M.G. CURRY & B.F. MARTIN. Upper Bear Creek Reservoirs. (Eugene M. 1975. A Vascular Flora of St. Helena and Futato, ed.). Reports of Investigations West Feliciana Parishes, Louisiana. Univ. No. 29, Office of Archaeological Research, Southwestern Louisiana, Biology Research Univ. Alabama. Series No. 39, Lafayette. _____. 1982b. Plant remains from the ANDERSON, E. 1952. Plants, Man and Life. Yarborough site. Pp. 134-140 in Archa­ Univ. California Press, Berkeley. eological Investigations at the Yarborough ASCH, D.L. and N.E. ASCH. 1985. Prehistoric Site (22CL814) (Carlos Solis and Richard plant cultivation in west-central illinois. Walling, eds.). Reports of Investigations Pp. 149-203 in Prehistoric Food Production No. 30, Office of Archaeological Research, in North America (Richard I. Ford, ed.). Univ. Alabama. Univ. Michigan Museum of Anthropology, _____. 1982c. Plant Resources, Archa- Anthropological Papers No. 75. eological Plant Remains and Prehistoric AVERY, G.N. and L.L. LOOPE. 1980. Plants Plant-use Patterns in the Central Tombig­ of Everglades National Park. Report T-574, bee River Valley. Alabama Museum of Everglades National Park, South Florida Natural History No.7, Univ. Alabama. Research Center, Homestead, Florida. _____. 1983. Floral remains from the BAKER, H.G. 1974. The evolution of weeds. Lubbub Creek archaeological locality. Ann. Rev. Eco!. Syst. 5:1-24. Pp. 1-78 in Prehistoric Agricultural Com­ BOWEN, W.R. 1982. Archaeological Investi­ munities in West Central Alabama. (C.S. gations at 9Ck(DOT)7, Cherokee County, Peebles, ed.). Volume 2, Studies of Georgia. Georgia Dept. Transportation, Material Remains from the Lubbub Creek Occasional Papers in Cultural Resource Archaeological Locality. U.S. Army Corps Management, No.1. of Engineers, Mobile District, National BRAUN, E.L. 1943. An Annotated Catalog of Technical Information Services, Spring­ Spermatophytes of Kentucky. John C. field, Virginia. Swift and Company, Cincinnati. CHAPMAN, J. and G. CRITES. 1987. Evidence BROWN, C.A. 1972. Wildflowers of Louisiana for early maize (Zea mays) from the Ice­ and Adjoining States. Louisiana State house Bottom site, Tennessee. Am. Antiq. Univ. Press, Baton Rouge. 52:352-354. BROWN, M.L. and R.G. BROWN. 1984. Herba­ CHAPMAN, J. and A.B. SHEA. 1981. The ceous Plants of Maryland. Port City Press, archaeobotanical record: Early Archaic Baltimore. period to contact in the lower Little Ten­ BYRD, K.M. and R.W. NEWMAN. 1978. Archa­ nessee River valley. Tennessee Anthro­ eological data relative to prehistoric sub­ pologist 6:61-84. sistence in the lower Mississippi alluvial CHAPMAN, J., P.A. DELCOURT, P.A. valley. Pp. 9-21 in Geoscience and Man CRIDLEBAUGH, A.B. SHEA and H.R. Volume XIX. Louisiana State Univ., Baton DELCOURT. 1982. Man-land interaction: Rouge. 10,000 years of American Indian impact CADDELL, G.M. 1982a. Plant remains from on native ecosystems in the lower Little the Cedar Creek and Upper Bear Creek Tennessee River valley. Southeastern Reservoirs. Pp. 335-349 in Archaeological Arch. 2:115-121. Investigations in the Cedar Creek and CLEWELL, A.F. 1985. Guide to the Vascular Winter 1989 JOURNAL OF ETHNOBIOLOGY 153

LITERATURE CITED (continued)

Plants of the Florida Panhandle. Florida GARDNER, P.S. 1982. Carbonized plant re­ State Univ. Press, Tallahassee. mains. Pp. 215-226 in Archaeological Data COTTIER, J. 1971. Towasahgy State Archa­ Recovery in the Richard B. Russell Multiple eological Site, Missouri. State Park Board, Resource Area: The Anderson County Jefferson City. Group (W.O. Wood, ed.). Interim Report CRAWFORD, G. 1975. Letter to Dr. Roy S. Contract No. C-54048(81). V.S. Army Dickens, Jr. dated 9/21/75. On file, Re­ Corps of Engineers, Columbia, South search Laboratories of Anth., Univ. North Carolina. Carolina! Chapel Hill. _____. 1984. Paleoethnobotany of the CUTLER, H.C. and L.W. BLAKE. 1976. Plants Beaverdam Creek mound and village site from Archaeological Sites East of the (9Eb85), Elbert County, Georgia. Pp. 485­ Rockies. American Arch. Reports No.1, 504 in Archaeological Investigations at Microfiche, American Archaeology Divi­ the Beaverdam Creek Site (9Eb85), Elbert sion, Univ. Missouri, Columbia. County, Georgia (by J.L. Rudolph and D.J. DEAM, C.C. 1940. Flora of Indiana. Dept. Hally). Submitted to the National Park Conservation, Div. Forestry, Indianapolis. Service. DEWAR, R. 1982. Dump heaps revisited. Ms. _____. 1985. Plant remains from Lamar in author's possession. Village, Georgia. MS. on file, Research DUNN, M.E. 1982. Botanical remains from Laboratories of Anth., University of North the Cemochechobee site. Pp. 252-255 in Carolina, Chapel Hill. Cemocheechobee, Archaeology of a Missis­ GREAT PLAINS FLORA ASSOCIATION. 1977. sippian Ceremonial Center on the Chat­ Atlas of the Flora of the Great Plains. tahoochee River. (By F.T. Schnell, V.J. Iowa State Univ. Press, Ames. Knight, Jr., and G.S. Schnell). Univ. GREMILLION, K.J. 1987a. Plant remains. Pp. Presses of Florida, Gainesville. 127-157 in The Historic Occaneechi: An ESPENSHADE, E.B. (ed.). 1982. Goode's World Archaeological Investigation of Culture Atlas. 16th edition. Rand McNally and Change, Final Report of 1986 Excavations Company, Chicago. (H.T. Ward and R.P.S. Davis Jr., eds.). FAULKNER, C.H. 1977. The Eoff I site (40CF32). Research Laboratories of Anth., Univ. Pp. 65-278 in Fourth Report of the Nor­ North Carolina, Chapel Hill. mandy Archaeological Project, Normandy _____. 1987b. Plant remains from the Archaeological Project Volume 4. (C.H. Fredricks, Wall, and Mitchum sites. Pp. Faulkner and Major C.R. McCollough, 259-277 in The Siouan Project: Seasons I eds.). Dept. Anthroplogy, Univ. Tennes­ and II (R.S. Dickens, Jr., H.T. Ward and see, Reports of Investigations No. 19. R.P.S. Davis Jr., eds.). Monograph Series Knoxville. No.1, Research Laboratories of Anth., FERNALD, M.L. 1950. Gray's Manual of Univ. North Carolina, Chapel Hill. Botany. Eighth edition. American Book HALLY, D.J. 1981. Plant preservation and the Company, New York. context of paleobotanical samples: a case FRITZ, G.J. 1982. Analysis of plant remains study. Amer. Antiq. 46:723-742. from the Spiro site, 34Lf46, LeFlore HARVILL, A.M., C.E. STEVENS and D.M.E. County, Oklahoma. Pp. 201-213 in Spiro WARE. 1977. Atlas of the Virginia Flora, Archaeology, 1980 Research (by J. Daniel Part II: Dicotyledons. Virginia Botanical Rogers). Oklahoma Arch. Survey Studies Assoc., Farmville. in Oklahoma's Past No.9, Norman. HEDRICK, V.P. (ed.). 1972. Sturtevant's Edible _____. 1984. Evidence of plant use Plants of the World. Dover Books, New from Copple Mound at the Spiro site. York. Report submitted to the Spiro Project, HUTCHINSON, J. 1967. The Genera of Flower­ Oklahoma Arch. Survey, Norman. ing Plants (Angiospermae). Volume II, _____. 1986. Prehistoric Ozark Agricul­ Dicotyledones. Clarendon Press, Oxford. ture: The University of Arkansas Rock­ KELLERMAN, W.A. and W.C. WERNER. 1894. shelter Collections. Ph.D. Dissertation Catalogue of Ohio Plants. Report of the (Anthroplogy), Vniv. North Carolina, Ohio Geol. Survey, Columbus. Chapel Hill. KILLIP, E.P. 1938. The American species of 154 GREMILLION Vol. 9, No.2

LITERATURE CITED (continued)

Passifloraceac. Botanical 5C'ries, Field SCARRY, C.M.M. 1980. Plant procurement Mus{'um of Natural History Volume XIX, strategies in the West Jefferson and Parts I and II, Chicago. Moundville I phases. Paper presented at _____.1960. Supplemental not{'son the the 37th Annual Meeting of the South­ American species of PassifJoraceae with eastern Archaeological Conference, New descriptions of new species. Pp. 1·23 in Orleans. Contributions from thc U.s. Nat. Herb­ SCARRY, C.M. 1986. Change in plant procure­ arium, Volume35. Smithsonian Institution ment and production strategies during the Press, Washington, DC. emergence of the Moundville chiefdom. LAKELA, O. 1965. Annotated Checklist of thc Ph.D. Diss('rt., (Anthropology), Univ, Vascular Plants of Collier, Dade, and Michigan. Ann Arbor. Monr(>e Counties. Florida. Univ. Miilmi SCHROEDL, G.F. and A.B. SHEA. 1986. Plant Press, Coral Gables. remains. Pp. 515-530 in Overhill Archa­ MAJOR, R.H., ed. 1849. The Historic of Tra­ eology at Chota-Tanassee (Gerald F. vaile intu Virginia Britannia, by William Schroedl, ed.). Univ. Tennessee, Dept. of Strachey. London. Anthropology, Report of Investigations MARTIN, F.W. and H,Y, NAKASONE. 1970. No. 28. The edible spedes of PIl5Sifiora. Econ. SCHROEDL, C.R., R.P.s. DAVIS JR. and Botany 24:333-343. c.c. BOYD JR. 1985. Archaeological Con­ MCMAHAN, J.D. 1983. Palcoethnobol

LITERATURE CITED (continued)

SMALL, J.K. 1913. Flora of Miami. Privately WATSON, P.]. 1985. The impact of early horti­ printed, New York. culture in the upland drainages of the SMITH, B.D. 1987. The independent domesti­ midwest and midsouth. Pp. 99-148 in cation of indigenous seed-bearing plants in Prehistoric Food Production in Eastern Eastern North America. Pp. 3-47 in Emer­ North America (ed. R.I. Ford). Univ. gent Horticultural Economies of the Michigan Museum of Anth., Anthro­ Eastern Woodlands (W.F. Keegan, ed.). pological Papers No. 75. Southern Illinois Univ. Center for Archa­ WHERRY, E.T., ].M. FOGG and H.A. WAHL. eological Investigations, Occasional Paper 1979. Atlas of the Flora of Pennsylvania. No.7, Carbondale. Morris Arboretum. SMITH, E.B. 1978. An Atlas and Annotated WILSON, J.H. ]R. 1977. Feature Fill, Plant List of the Vascular Plants of Arkansas (and Utilization and Disposal Among the Supplements). Privately printed. Historic Sara Indians. Master's thesis STEYERMARK, J. 1962. Flora of Missouri. Iowa (Anthropology) Univ. North Carolina, State Univ. Press, Ames. Chapel Hill. STRAUSBAUGH, P.D. and E.L. CORE. 1978. WRIGHT, L.B., ed. 1947. (orig. 1705). The Flora of West Virginia. Seneca Books, Inc., History and Present State of Virginia (by Grantsville, West Virginia. Robert Beverley). Univ. North Carolina TRINKLEY, M. 1976. Ethnobotanical analysis Press, Chapel Hill. of selected material from 44Ha22. Report YARNELL, R.A. 1987. A survey of prehistoric submitted to Howard A. MacCord, Ms. crop plants in eastern North America. on file. Research Laboratories of Anth. The Missouri Archaeologist, in press. Ethnobotanical Laboratory, Univ. North YARNELL, R. and M.J. BLACK. 1985. Temporal Caf{~lina, Chapel Hill. trends indicated by a survey of Archaic VANKAT, J.L. 1979. The Natural Vegetation and Woodland plant food remains from of North America: An Introduction. John southeastern North America. Southeastern Wiley and Sons, New York. Archaeology 4:93-106.

White Gold: The Diary of a Rubber Cutter in the Amazon, 1906-1916. John C. Yungjohann (edited by Ghillean T. Prance). Synergetic Press, POB 689, Oracle, Arizona. 1989. Pp. 103. Paperback only. $7.95 + 10% postage and handling.

How fortunate we are that John C. Yungjohann, a tile setter and not a trained naturalist, had the energy to keep a diary during his years as a rubber cutter in the Amazon region. Many, in fact most, of his fellow rubber cutters did not survive the trails and tribulations, the adventures and misadventures which the author not only experienced but also wrote about. And he writes of joyous moments too. His account of how he came to make friends with one tribe of local Indians shows his determination and ingenuity. This provided him an opportunity to observe them, their life styles, activities and uses of plants and animals. This, along with the many descriptions of the plants and animals he encountered in one way or another, will be of interest to readers of this journal. Since the author used local common names for these plants and animals, Editor Ghillean T. Prance has provided informative commentary and photography to identify the plants and animals and interpret their uses.